Just Transit Transit Dependents Civil Rights and Transit

  • Slides: 75
Download presentation
Just Transit? Transit Dependents, Civil Rights, and Transit Policy Brian D. Taylor Hiroyuki Iseki

Just Transit? Transit Dependents, Civil Rights, and Transit Policy Brian D. Taylor Hiroyuki Iseki UCLA Institute of Transportation Studies October 2001 Lake Arrowhead, California

Transit Equity? • In recent years, most equity debates in public transit concerned the

Transit Equity? • In recent years, most equity debates in public transit concerned the “fair” distribution of resources among jurisdictions. • Tendency is toward treating voters equally.

Transit Equity? • But the use of transit is very uneven spatially. ØThe 10

Transit Equity? • But the use of transit is very uneven spatially. ØThe 10 largest U. S. transit systems account for over 60% of all transit trips. ØAbout 1/3 of all U. S. transit trips are taken in metropolitan New York.

Transit Equity? • The results are uneven in two ways: 1. Transit systems in

Transit Equity? • The results are uneven in two ways: 1. Transit systems in the oldest and largest central cities receive the highest taxpayer subsidies, in absolute terms. 2. Transit riders on newer, smaller suburban transit systems tend to receive the highest taxpayer subsidies, in relative terms.

Transit Equity?

Transit Equity?

Why do we subsidize public transit? • Direct benefits: Provide mobility for those without

Why do we subsidize public transit? • Direct benefits: Provide mobility for those without access to private vehicles, and travel options for those who choose not to drive. • Indirect benefits: Decrease traffic congestion and reduce travel times for all travelers; reduce energy consumption, vehicle emissions, suburban sprawl, amount of land devoted to roads, motor vehicle noise, and accident costs. • Network/Service economies: Subsidies are required to maintain comprehensive transit route networks with sufficient service frequencies to avoid a downward spiral of declining ridership and service.

Why do we subsidize public transit? • Compensate for policy bias in favor of

Why do we subsidize public transit? • Compensate for policy bias in favor of private vehicles: Subsidize transit fares to compensate for public policies which do not fully charge drivers for the social costs of auto use. • Equity: Public transit is an indispensable social service that provides access to basic needs for transit dependents.

The Truth About Transit: Most Transit Users are Bus Riders, and Most Bus Riders

The Truth About Transit: Most Transit Users are Bus Riders, and Most Bus Riders are Poor Source: 1995 Nationwide Personal Transportation Survey. Overall, 1/3 of all transit users come from households with 1995 incomes below $15, 000, and 3/5 from households with 1995 incomes below $30, 000 (Pucher, 1998).

Conflicting Policy Goals Serve low-income transit dependents Expand commuteroriented services ISTEA and TEA-21 mandates

Conflicting Policy Goals Serve low-income transit dependents Expand commuteroriented services ISTEA and TEA-21 mandates require the expenditure of federal funds for transportation purposes to comply with Title VI of the Civil Rights Act; these have prompted civil rights lawsuits against transit agencies over supposedly discriminatory fare policies and/or expansion of commuteroriented services.

Policy Questions Guiding This Issue • Who “wins” and who “loses” in the subsidy

Policy Questions Guiding This Issue • Who “wins” and who “loses” in the subsidy of public transit? – In general, is the subsidy of public transit progressive or regressive with respect to income? – More specifically, how are transit subsidies distributed among various classes of riders?

Findings in a Nutshell • Because the tax system is generally progressive, transit subsidies

Findings in a Nutshell • Because the tax system is generally progressive, transit subsidies tend to transfer benefits from higher-income people to low-income people.

Findings in a Nutshell • Because the tax system is generally progressive, transit subsidies

Findings in a Nutshell • Because the tax system is generally progressive, transit subsidies tend to transfer benefits from higher-income people to low-income people. Ø although low-income non-transit-users are significant losers in this transfer.

Findings in a Nutshell 1. Because the tax system is generally progressive, transit subsidies

Findings in a Nutshell 1. Because the tax system is generally progressive, transit subsidies tend to transfer benefits from higher-income people to low-income people. Ø although low-income non-transit-users are significant losers in this transfer. Ø If the goal is simply income redistribution, there are better ways to do it than subsidizing transit.

Findings in a Nutshell 2. Among transit users, the distribution of transit subsidies is

Findings in a Nutshell 2. Among transit users, the distribution of transit subsidies is generally regressive with respect to income. Ø higher-income transit users tend to be subsidized more than low-income transit users. Ø Estimating this distribution of transit subsidies among various transit users, however, is a complex endeavor.

Analysis Flowchart

Analysis Flowchart

Data Sources • LA MTA Reports 1. Line Performance Trends Report 2. Consolidated Transit

Data Sources • LA MTA Reports 1. Line Performance Trends Report 2. Consolidated Transit Service Reports (Ridecheck Report) 3. Schedule Quality Report • National Transit Database

The Problem of Peaking The marginal cost of transit service is typically highest in

The Problem of Peaking The marginal cost of transit service is typically highest in the peak period and peak direction. • Lower labor efficiency: – Limits on the use of part-time labor and of split- and spread-time shifts. • Lower equipment utilization efficiency: – Extra vehicles needed to meet peak period demand, – A higher proportion of non-revenue service (extra deadheading and trippers), – Scaling facilities to accommodate peak service levels.

Time-of-Day Variation in Service Levels: Los Angeles MTA

Time-of-Day Variation in Service Levels: Los Angeles MTA

Comparison of Individual Line Costs Using a Partial Cost Allocation Model

Comparison of Individual Line Costs Using a Partial Cost Allocation Model

Comparison of Estimated Costs between then-current LA MTA Model and a Full-Cost Allocation Model

Comparison of Estimated Costs between then-current LA MTA Model and a Full-Cost Allocation Model

Peak/Base Ratios of the Twenty-Seven Largest Transit Operators

Peak/Base Ratios of the Twenty-Seven Largest Transit Operators

Comparison of Estimated Bus System and Light Rail Costs (Cost per Passenger Capacity Hour)

Comparison of Estimated Bus System and Light Rail Costs (Cost per Passenger Capacity Hour) Using MTA Model and Fully-Allocated Model

LA MTA Bus Rider Demographics Data source: Service Planning Market Research Program, FY 96

LA MTA Bus Rider Demographics Data source: Service Planning Market Research Program, FY 96 -97 MTA Bus On-Board Passenger Survey.

Per Trip Subsidies for Service Types by Income

Per Trip Subsidies for Service Types by Income

Per Trip Subsidies for Service Types by Race/Ethnicity

Per Trip Subsidies for Service Types by Race/Ethnicity

Per Trip Subsidies for Service Types by Age Group

Per Trip Subsidies for Service Types by Age Group

Per Trip Subsidies for Service Types by Sex

Per Trip Subsidies for Service Types by Sex

Per Trip Subsidies by Time of Day by Income

Per Trip Subsidies by Time of Day by Income

Factors Influencing Demographic Variation in Subsidies 1. Demographic variation in transit subsidies are due

Factors Influencing Demographic Variation in Subsidies 1. Demographic variation in transit subsidies are due to demographic variation in service consumption. Ø Most of the differences in subsidy levels by income are due (primarily) to the longer average trip distances of higher-income riders and (secondarily) to their greater use of capital-intensive and commuter-oriented modes.

Factors Influencing Demographic Variation in Subsidies 2. Subsidies by time of day vary little

Factors Influencing Demographic Variation in Subsidies 2. Subsidies by time of day vary little after controlling for trip distance and travel mode. Ø The higher unit costs of service supplied during peak periods are mitigated by higher levels of peak-period utilization in the case of the LA MTA. Ø This would likely change, however, should the MTA expand peak-period service in an effort to reduce standees.

Comparison of Estimated Bus System and Light Rail Costs (Cost per Seat Hour) Using

Comparison of Estimated Bus System and Light Rail Costs (Cost per Seat Hour) Using MTA Model and Fully-Allocated Model

Two Caveats. . . 1 The results presented here are preliminary. 2 While the

Two Caveats. . . 1 The results presented here are preliminary. 2 While the data are from Los Angeles, our focus is not on the MTA per se. • Given both the diversity of MTA riders and the low levels of peaking on MTA buses and trains, • the MTA probably has less demographic variation in subsidies than at most other transit operators nationwide.

Conclusions Because the subsidy of a transit trip is a function of the variable

Conclusions Because the subsidy of a transit trip is a function of the variable cost of that trip minus the fare paid by a traveler, the key to equalizing subsidies is a fare policy.

Conclusions 1. Transit managers should have a clear sense of the variability of service

Conclusions 1. Transit managers should have a clear sense of the variability of service production costs. 2. As a general principle, fares should be set to vary in rough proportion with costs. 3. The adoption of such a marginal, costbased fare structure can simultaneously increase both efficiency and equity in the use and subsidy of transit service

Conclusions Linking fares to costs… Ø Efficiency would increase by: 1. Encouraging passengers to

Conclusions Linking fares to costs… Ø Efficiency would increase by: 1. Encouraging passengers to consume more inexpensive-to-provide transit service (short, off-peak bus trips); 2. Encouraging passengers to be more judicious in their consumption of expensive-to-provide transit service (long, peak trips on capital intensive modes). In concert, these two factors would work to decrease overall subsidies per rider.

Conclusions Linking fares to costs… Equity would increase by: 1. Lowering (in relative terms)

Conclusions Linking fares to costs… Equity would increase by: 1. Lowering (in relative terms) the price of transit services disproportionately consumed by low-income passengers; and 2. Increasing (relatively) the price of transit services disproportionately consumed by higher-income passengers. In concert, these two factors would work to eliminate the regressivity of current transit subsidies.

Just Transit? Transit Dependents, Civil Rights, and Transit Policy Brian D. Taylor Hiroyuki Iseki

Just Transit? Transit Dependents, Civil Rights, and Transit Policy Brian D. Taylor Hiroyuki Iseki UCLA Institute of Transportation Studies October 2001 Lake Arrowhead, California

Conclusions 1. Transit managers should have a clear sense of the variability of service

Conclusions 1. Transit managers should have a clear sense of the variability of service productions costs. 2. As a general principle, fares should be set to vary in rough proportion with costs. 3. The adoption of such a marginal cost-based fare structure can simultaneously increase both efficiency and equity in the use and subsidy of transit service

Intended to be blank.

Intended to be blank.

Purpose of This Study • Develop a cost estimation method that, in contrast to

Purpose of This Study • Develop a cost estimation method that, in contrast to typical cost allocation models: 1. is more sensitive to cost variation by time of day, 2. takes into account vehicle and non-vehicle capital costs, takes into account the passenger capacity of vehicles in various transit modes. 3.

Data Sources • LA MTA Reports 1. Line Performance Trends Report 2. Consolidated Transit

Data Sources • LA MTA Reports 1. Line Performance Trends Report 2. Consolidated Transit Service Reports (Ridecheck Report) 3. Schedule Quality Report • National Transit Database

Transit Cost Allocation Models • Service costs are a function of service outputs. •

Transit Cost Allocation Models • Service costs are a function of service outputs. • Service outputs are most commonly measured in terms of: vehicle-hours, vehicle miles, and peak vehicles.

Partially- and Fully-Allocated Models • Partially-allocated model includes only variable costs and some semi-fixed

Partially- and Fully-Allocated Models • Partially-allocated model includes only variable costs and some semi-fixed costs to reflect the marginal costs of incremental service modifications. • Fully-allocated model includes most or all fixed costs for the use of comparing performance between modes or systems.

Average Cost Approach to Allocating Costs by Service Levels a, b: the number of

Average Cost Approach to Allocating Costs by Service Levels a, b: the number of required buses CP, CB : vehicle hour-related costs of the Peak and Base service U : unit cost of service output S : the ratio of peak costs to base costs Source: Adapted from Cervero (1980)

The General Form of a Cost Allocation Model C : estimated costs i :

The General Form of a Cost Allocation Model C : estimated costs i : a particular measurable service output which represents the scale of operations n : number of service outputs included in the model Ui : unit cost of service output i Xi : quantity or value of service output i in the analysis

Labor Utilization Factor for Vehicle Hour Unit Cost Source: Adapted from Yu (1986) and

Labor Utilization Factor for Vehicle Hour Unit Cost Source: Adapted from Yu (1986) and Cherwony and Mundle (1978, 1980) • • LUFi : Labor Utilization Factor for period i PHi : pay hours for period i VHi : vehicle hours for period i n : relative labor productivity (= (PHP/ VHP) / (PHB /VHB) = 1. 302) s : vehicle hour coefficient (= VHP / VHB) PHP or B : pay hours for peak or base period VHP or B : vehicle hours for peak or base period

Accounting for Vehicle Utilization • Vehicle Utilization: Lower during peak periods – To account

Accounting for Vehicle Utilization • Vehicle Utilization: Lower during peak periods – To account for the variability in deadheading and inter-lining between time periods, expenses were allocated to each period on the basis of total (or “scheduled”) vehicle miles, but costs were calculated on the basis of revenue (or “in-service”) vehicle miles and hours. Period Rev VM Tot VM Cost per revenue VM. Peak 30 miles 60 miles ==> $ 600 $600 / 30 miles = $20. 00 per revenue VM Off-peak 30 miles 40 miles ==> $ 400 $400 / 30 miles = $13. 33 per revenue VM Daily Total Cost $1, 000 Daily $1000 / 100 = $10 per SVM Peak 60 / 30 = 2. 000 Off-peak 40 / 30 = 1. 333 $10 * 2. 000 = $20. 00 $10 * 1. 333 = $13. 33

Including Semi-Fixed and Vehicle Capital Costs C 1 : The cost assigned to the

Including Semi-Fixed and Vehicle Capital Costs C 1 : The cost assigned to the base period C 2 : The costs of the additional peak service Cp : The costs in the Peak period (2 t 2) U : unit cost of service output in each period Source: Adapted from Levinson (1978) and Cervero (1980)

Marginal Cost Approach to Allocating Vehicle Capital Costs Assumption: All buses in service during

Marginal Cost Approach to Allocating Vehicle Capital Costs Assumption: All buses in service during base periods are available for use during peak periods.

Modified Cost Models • Fully-Allocated Model -- Including all variable, semi-fixed, and fixed costs

Modified Cost Models • Fully-Allocated Model -- Including all variable, semi-fixed, and fixed costs FACi, j = OCi, j + CCi, j = ( LUFi, j * UVH * VHi, j + UVM * VMi, j + PVCi, j + UTP * TPi, j ) * ( 1 + F ) + VCCi, j + OCC * ( IVHi, j / IVHday, system) ( LUFi, j = 1, F = UTP = 0 for LRT ) • Partially-Allocated Model I -- Excluding non-vehicle capital costs PACi, j = OCi, j + VCCi, j = ( LUFi, j * UVH * VHi, j + UVM * VMi, j + PVCi, j + UTP * TPi, j ) * ( 1 + F ) + VCCi, j • Partially-Allocated Model II -- Including only variable and vehicle capital costs PACi, j = OCi, j + VCCi, j = ( LUFi, j * UVH * VHi, j + UVM * VMi, j ) + VCCi, j

Comparison of Individual Line Costs Using Partially-Allocated Model I

Comparison of Individual Line Costs Using Partially-Allocated Model I

Cost of Additional Vehicle Run for Five Sample Bus Lines by Time Periods

Cost of Additional Vehicle Run for Five Sample Bus Lines by Time Periods

Findings-Phase I In this case study comparing these models with the one currently used

Findings-Phase I In this case study comparing these models with the one currently used by the LA MTA: 1. Peak period bus costs are estimated to be higher (36 %); 2. Base period bus costs are estimated to be lower (15 %); 3. LRT unit costs are estimated to be substantially higher than bus costs (57~113 %) per unit of passenger capacity. 4. Incremental changes in bus service are estimated to be substantially lower (29~75 %), regardless of time-of-day.

Linking Model to Travel Data • Travel demographic data are drawn from the annual

Linking Model to Travel Data • Travel demographic data are drawn from the annual MTA On-board Passenger Survey. • Survey gathers information on trip characteristics and travel demographics. • Approximately 5, 000 surveys are collected each year; our cleaned, 3 -year data set contains 10, 710 records. • The cost per passenger mile estimated from our model was applied to each record. • These estimated trip costs can then be aggregated to compare differences among different demographic groups.

Preliminary Findings-Phase II • Consistent with the findings of other travel behavior surveys, the

Preliminary Findings-Phase II • Consistent with the findings of other travel behavior surveys, the patterns of transit travel in Los Angeles vary systematically by rider demographics. • The transit trips consumed by higher-income riders, men, and whites are estimated to be, on average, more expensive than the trips consumed by lower-income riders, women, and non-whites. • These differences are mitigated somewhat, though not entirely, by the higher-fares charged for longer-distance express and transfer trips.

Conclusions • Current models do not account for the high degree of variability in

Conclusions • Current models do not account for the high degree of variability in the cost of providing transit service. • Such shortcomings make it difficult decision makers to make informed decisions regarding expansions, changes, or deletions of service. • The preliminary results of our demographic analysis suggest that current patterns of transit utilization combine with relatively flat fare policies to favor higher-income riders over lower-income riders. • This preliminary analysis supports the linking of transit fares to costs on both efficiency and equity grounds.

Just Transit? Transit Dependents, Civil Rights, and Transit Policy

Just Transit? Transit Dependents, Civil Rights, and Transit Policy

Limitations of Current Cost Allocation Models • The models typically used in practice include

Limitations of Current Cost Allocation Models • The models typically used in practice include all operating costs and exclude all capital costs, without regard to how such costs vary with the provision of service. • The models are highly aggregated; they are typically based on systemwide costs and do not account for cost variation on individual routes, between various modes, or by time of day.

Next Steps • Accounting for the directional peaking of demand, • Taking weekend operation

Next Steps • Accounting for the directional peaking of demand, • Taking weekend operation directly into account in computing vehicle and capital costs, • Applying a “cost centers” approach to differentiate unit costs to discrete parts of the system, • Computing relative labor productivity factors on individual lines from the ratio of pay hours to vehicle hours by time of day.

Figures for Cost Comparison Charts Comparison of Estimated Systemwide Costs between the MTA Model

Figures for Cost Comparison Charts Comparison of Estimated Systemwide Costs between the MTA Model and the Fully-Allocated Model Comparison of Estimated Bus System and LTR Costs

Summary of Modification • Modify MTA’s cost allocation model to more accurately estimate costs

Summary of Modification • Modify MTA’s cost allocation model to more accurately estimate costs on individual lines in different periods of day for the Metro Bus system – Labor Utilization Factor for VH costs – Vehicle Usage Apportionment Factor (VUAF) and Weekend Operation Factor (WOF) for PV costs – Proportional allocation of VM and TP costs • Include capital costs to compare system wide costs between the Metro Bus System and the Blue Line – VUAF and WUF for vehicle capital costs – Equal distribution of other capital costs by VH • Use of In-service VH to compute costs in comparison to account for labor and vehicle utilization efficiency

Blue Line Operating Cost Allocation Unit Cost Calculation

Blue Line Operating Cost Allocation Unit Cost Calculation

MTA Cost Allocation Model OCj = ( UVH * VHj + UVM * VMj

MTA Cost Allocation Model OCj = ( UVH * VHj + UVM * VMj + UPV * PVj + UTP * TPj ) *(1+F) • • OC j U VH VM PV TP F : estimated operating costs : unit of analysis in question — system, line, etc. : unit cost per service output : scheduled vehicle hours : scheduled vehicle miles : PM peak vehicles : total passengers : fixed overhead cost factor

Difficulties in the Study • MTA’s different reports used different definition of time periods.

Difficulties in the Study • MTA’s different reports used different definition of time periods. For example, the AM peak is 5 -9 a. m. in one report, while it is 6 -9 a. m. in another. • Data were obtained in different years for different reports. • Many data such as SVH, SVM, IVH, IVM, PV, and TP were not available for all six periods of day. (apportionment) • IVH and IVM data were not available for some bus lines. (regression) • Details of expense item assignment in MTA’s model for the Metro bus system were unknown. • Nor was there a detailed breakdown of operating costs for the Blue Line. • There is no specific way to compute annual capital costs.

Marginal Cost Approach to Allocating Vehicle Capital Costs Assumption: All buses in service during

Marginal Cost Approach to Allocating Vehicle Capital Costs Assumption: All buses in service during the period with the smaller number of in-service vehicles will be utilized in any other periods that have higher vehicle requirements.

PREMISE: The public finance of transportation is guided first and foremost by concerns over

PREMISE: The public finance of transportation is guided first and foremost by concerns over equity

Unjust Equity • What one individual, group, organization, or jurisdiction views as fair and

Unjust Equity • What one individual, group, organization, or jurisdiction views as fair and equitable, another may view as unfair and unjust. • Both views may be correct.

Types of Equity • Market Equity: Bring prices in line with costs imposed and/or

Types of Equity • Market Equity: Bring prices in line with costs imposed and/or benefits received • Opportunity Equity: Treat individuals, interest groups, or jurisdictions equally • Outcome Equity: Redistribute resources to effect equal outcomes

Why do people debating equity in transportation seem so often to be talking past

Why do people debating equity in transportation seem so often to be talking past one another?

Because they focus on different units of analysis

Because they focus on different units of analysis

Units of Analysis in Transportation Policy • Individuals: residents, voters, travelers, etc. • Groups:

Units of Analysis in Transportation Policy • Individuals: residents, voters, travelers, etc. • Groups: modal interests, racial/ethnic groups, etc. • Areas (geographic): states, counties, legislative districts, etc.

Factors Influencing Demographic Variation in Subsidies 1. Demographic variation in transit subsidies are due

Factors Influencing Demographic Variation in Subsidies 1. Demographic variation in transit subsidies are due to demographic variation in service consumption. Most of the differences in subsidy levels by income are due (primarily) to the longer average trip distances of higher-income riders and (secondarily) to their greater use of capital-intensive and commuter-oriented modes. 2. Subsidies by time of day vary little after controlling for trip distance and travel mode. The higher unit costs of service supplied during peak periods are mitigated by higher levels of peak period utilization in the case of the LA MTA. This would likely change, however, should the MTA expand peak period service in an effort to reduce standees.

Transit Equity? • The results are uneven in two ways: 1. Transit systems in

Transit Equity? • The results are uneven in two ways: 1. Transit systems in the oldest and largest central cities receive the highest taxpayer subsidies, in absolute terms. 2. Transit riders on newer, smaller suburban transit systems tend to receive the highest taxpayer subsidies, in relative terms.